Table 5 Performance Comparison of MLP and other Attention Modules.

From: A simple monocular depth estimation network for balancing complexity and accuracy

Method

\(\delta _1\uparrow\)

AbsRel\(\downarrow\)

RMSE\(\downarrow\)

Params\(\downarrow\)

Encoder(MSCAN)39+Decoder(Ours)+MLP

0.911

0.097

0.348

30.9M

Encoder(MSCAN)39+Decoder(Ours)+mViT12

0.912

0.097

0.349

35.5M

Encoder(MSCAN)39+Decoder(Ours)+MSA35

0.914

0.096

0.349

30.3M

Encoder(MSCAN)39+Decoder(Ours)+WAT

0.916

0.095

0.345

30.6M

  1. The best result is indicated in bold.
Back to article page